Lightweight Attribute Localizing Models for Pedestrian Attribute Recognition

TL;DR

This paper introduces a novel low-rank compression method for pedestrian attribute recognition models that preserves gradient directions, enabling efficient model size reduction without significant accuracy loss.

Contribution

It proposes an optimal rank determination approach for low-rank layers that maintains gradient alignment, improving compression efficiency for PAR models.

Findings

Significant reduction in model complexity achieved

Maintains high accuracy after compression

Effective for resource-constrained deployment

Abstract

Pedestrian Attribute Recognition (PAR) focuses on identifying various attributes in pedestrian images, with key applications in person retrieval, suspect re-identification, and soft biometrics. However, Deep Neural Networks (DNNs) for PAR often suffer from over-parameterization and high computational complexity, making them unsuitable for resource-constrained devices. Traditional tensor-based compression methods typically factorize layers without adequately preserving the gradient direction during compression, leading to inefficient compression and a significant accuracy loss. In this work, we propose a novel approach for determining the optimal ranks of low-rank layers, ensuring that the gradient direction of the compressed model closely aligns with that of the original model. This means that the compressed model effectively preserves the update direction of the full model, enabling more efficient compression for PAR tasks. The proposed procedure optimizes the compression ranks for each layer within the ALM model, followed by compression using CPD-EPC or truncated SVD. This results in a reduction in model complexity while maintaining high performance.